In: Computer Science
Why use parallelism? In a 500-word paper, define the
ways of measuring speed of computers, and discuss how parallelism
is used in super- computing. Include in your answer:
The ways we can measure speed.
Examples of super-computing and how it is being
used.
Answer)
Parallel programming is a set of defined program which get to be designed for considering a faster execution instead of sequential counterpart that would bring in a market advantage and in some of the other cases it is the speed which is being used for programs and in many of the cases, there is a faster execution which is much better and equated better.
The ways for measuring the speed of computers mostly measure the megahertz or that of gigahertz where one million ticks per second would be with one.
Having a clock speed for the computer would mostly measure the Gigahertz or the Megahertz. One can use the clock speed with regards to a rough measurement for a way the fast computer works.
The parallel computer would mostly be classified as per the level where the hardware would support the parallelism with that of the multi-core or the multiprocessor computer for having many processing elements with that of a single machine while clustering, griding and using multiple computers for working on a similar task.
Example: Supercomputers are special-purpose computers which have Belle, Hydra, and some like Gravity Pipe for astrophysics, MDGRAPE-3 for a protein structure computation molecular dynamics, or deep cracks for removing or breaking the DES Cipher.
Mostly a modern supercomputer would work more quickly with regards to splitting the problem as a piece and then working on most of the pieces which are dealt with parallel processing and
The computer speed is measured with some of the aspects like the clock speed where the signal remains constant, having a comparison of the processes and machines which would execute the operations and here the system will not indicate as to how quickly the system will be performed with regards to the system.
Some of the kind of forms for that of the parallel computing includes the bit level, instruction level, task or data parallelism which would have a high performance computing and fetches the broader interest because of the physical constraint for preventing the frequency scaling.
Mostly the parallel computer are being classified as a level wherein the hardware being supported by the parallelism while this could be multi core and multi processor computer that has multi processing element with that of single machine that has clusters, grids and MPPS that use the multi computer for working on a similar task and are mostly specialized in parallel computer architecture that would use specific processors for specific tasks.
In certain case of the parallism this would have a transparent for a programmer like the bit level or the instruction level parallelism while this would have an explicit parallel algorithm that would use the concurrent and could be difficult in writing the sequential one as the concurrency would introduce many classes for potential software bug that would have a race condition where synchronization and communication are a part.
Please
comment if you need any other inputs.
Share a like if you find the answer helpful. This would help us
learn and share our learning better.
Thank you.